fix: prune row groups correctly for columns with the same name

[evenyag]

I hereby agree to the terms of the GreptimeDB CLA.

Refer to a related PR or issue link (optional)

What's changed and what's your intention?

This PR fixes an issue that the parquet reader might prune row groups unexpectedly if two different columns are using the same name. This is possible if we drop a column and then add a column with the same name.

This PR passes the metadata of the region to the RowGroupPruningStats so it can get the correct column id to read.

For example, if we have the following table:

CREATE TABLE test(i TIMESTAMP TIME INDEX, j INTEGER, k INTEGER NOT NULL);
INSERT INTO test(i, j, k) VALUES (1, 11, 5), (2, 12, 5);
SELECT FLUSH_TABLE('test');
ALTER TABLE test DROP COLUMN j;
ALTER TABLE test ADD COLUMN j INTEGER DEFAULT 0;
INSERT INTO test(i, j, k) VALUES (3, 0, 6);
INSERT INTO test VALUES (4, 7, 0);

Before this PR:

SELECT * FROM test order by i;
+-------------------------+---+---+
| i                       | k | j |
+-------------------------+---+---+
| 1970-01-01T00:00:00.001 | 5 | 0 |
| 1970-01-01T00:00:00.002 | 5 | 0 |
| 1970-01-01T00:00:00.003 | 6 | 0 |
| 1970-01-01T00:00:00.004 | 7 | 0 |
+-------------------------+---+---+

SELECT * FROM test WHERE j = 0 order by i;
+-------------------------+---+---+
| i                       | k | j |
+-------------------------+---+---+
| 1970-01-01T00:00:00.003 | 6 | 0 |
| 1970-01-01T00:00:00.004 | 7 | 0 |
+-------------------------+---+---+

After this PR:

SELECT * FROM test WHERE j = 0 order by i;
+-------------------------+---+---+
| i                       | k | j |
+-------------------------+---+---+
| 1970-01-01T00:00:00.001 | 5 | 0 |
| 1970-01-01T00:00:00.002 | 5 | 0 |
| 1970-01-01T00:00:00.003 | 6 | 0 |
| 1970-01-01T00:00:00.004 | 7 | 0 |
+-------------------------+---+---+

Checklist

• I have written the necessary rustdoc comments.
• I have added the necessary unit tests and integration tests.
• This PR does not require documentation updates.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 85.30%. Comparing base (2d0f493) to head (5f58c33).
Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3802      +/-   ##
==========================================
- Coverage   85.62%   85.30%   -0.32%     
==========================================
  Files         947      953       +6     
  Lines      161925   162886     +961     
==========================================
+ Hits       138647   138957     +310     
- Misses      23278    23929     +651     

[fengjiachun]

LGTM

[waynexia]

LGTM